A Systematic Review and Meta-Analysis of the Timing of Vasopressor Therapy in Patients with Septic Shock: Assessing Clinical Outcomes and Implication

Information sources and searches

PubMed, Medline, Cochrane Library, Web of Science, and Google Scholar databases were comprehensively searched for potential studies published from inception until October 2023. Moreover, the bibliographies of articles from these databases were scrutinized for additional studies. Articles were retrieved from the electronic databases if they had the following MeSH terms and keywords: (Timing OR timely OR delay OR early) AND (vasopressors OR vasopressor therapy OR norepinephrine OR vasopressin OR vasoconstrictor agents OR epinephrine OR dopamine) AND (Septic shock). Furthermore, to avoid publication bias and improve our scientific research, we did not retrieve grey literature.

Eligibility criteria

Two independent reviewers analyzed the potential studies according to the predefined eligibility criteria. Articles were included in the present review if they were non-randomized or randomized studies with full articles and published in English, included adult patients with septic shock, and compared any vasopressor therapy according to the timing of administration. On the other hand, research articles that did not meet these criteria or were designed as reviews, conference abstracts, case reports, and editorials were excluded. Any discrepancy during this process was resolved via constructive debates between the two reviewers.

Data extraction and definitions

The data used for review and analysis was abstracted into separate Excel files by two independent reviewers. The differences in extracted data was initially resolved via constructive debates between the two reviewers, and if they failed to agree, an additional reviewer was engaged. The data abstracted included Author ID (surname of the primary author and date of publication), Study design, Study location, pertinent characteristics of the included patients (sample size, distribution of sex, and the number of patients receiving early or delayed vasopressor therapy), Vasopressor agents used, timing of vasopressor administration and outcomes.

The key endpoints of the current study were mortality rate and the duration needed to attain the goal mean arterial pressure (MAP) (≥65 mmHg), whereas the supplementary endpoints were the duration spent in ICUs and the incidence of new-onset arrhythmias. Mortality encompassed 28-day, 30-day, and in-hospital death rates.

Quality appraisal

The present review included randomized and non-randomized studies; therefore, assessment of methodological quality and risk of bias was performed using the Newcastle Ottawa Scale (NOS) and Cochrane’s risk of bias (RoB) tool, respectively. Using the NOS, articles were assessed according to the selection, comparability, and outcome domains. For each domain, a maximum of one star was assigned to each criterion fully addressed; otherwise, no star was assigned. For the selection domain, a maximum of four stars could be achieved, while the maximum number of stars that could be attained under the comparability and outcome domains were 2 and 3, respectively.

On the other hand, risk of bias assessment was performed according to the risk of selection, attrition, performance, reporting, and other biases. A low risk of bias was assigned to a domain that was sufficiently addressed within the study, while a high and unclear risk was assigned to domains not addressed entirely or with insufficient information to make a judgment.

Data synthesis

All quantitative statistical analyses in the present systematic review were conducted using the Review Manager software (RevMan 5.4.1). These analyses were pooled together using the DerSimonian and Laird random-effect model, which provides conservative effect sizes and counteracts the interstudy heterogeneity. Moreover, the dichotomous outcomes, such as mortality and incidences of new-onset arrhythmias, were pooled together using the simple odds ratio (OR), while continuous outcomes, such as ICU stay and time taken to achieve target MAP (≥65 mmHg), were pooled using the mean difference (MD) calculations. For studies in which the continuous outcomes were presented as median, interquartile ranges, and ranges, the formula provided by Hozo and colleagues was used to calculate the mean and standard deviations.¹⁸ In addition, the 95% confidence interval was employed for calculations, and a p-value of less than 0.05 was considered significant. Interstudy heterogeneity was calculated using the I2 statistics, of which values greater than 50% were considered high or significant. During the analyses, we grouped the outcomes according to the type of vasopressor therapy used, and whenever possible, subgroup analyses were performed.

Study selection

With the predefined keywords and MeSH terms, 2021 potential studies were identified from the electronic databases. A review of these articles led to the exclusion of 989 close or exact duplicates. Afterward, the remaining studies were screened based on their titles and abstracts, of which 917 deemed to be unrelated to our topic were excluded. Eventually, 12 studies were deemed eligible for review and analysis, while the remaining 64 articles were excluded as follows: 8 were animal studies, 12 did not have comparative groups, 4 included pediatric patients, 7 included patients with other conditions such as hemorrhagic shock and sepsis without evidence of shock, and 26 compared various vasopressin without showing the effect of the time of administration (Figure 1).

Figure 1. PRISMA flow chart.

Summary of study characteristics

Twelve studies with 2721 septic shock patients were included in the present review. Four of these studies were randomized controlled trials (RCT), and 8 were cohort studies. Furthermore, seven studies evaluated the timing of norepinephrine, and 5 assessed the timing of vasopressin as an adjunct treatment (Table 1).

Quality appraisal outcomes

The risk of bias outcomes are summarized in Figure 2. The bias assessment showed that most of the studies had a high risk of other bias because they were performed in single institutions. On the other hand, methodological quality of cohort studies is presented in Table 2. Most of the studies were carried out in single institutions; therefore, the could not attain maximum stars in the selection domain. Furthermore, two of the studies did not attain any stars in the comparability domain because they did not adjust any cofounders.

Figure 2. Risk of bias summary.

Does the timing of Norepinephrine administration improve outcomes of septic shock?

Of the 12 studies, 7, including 1678 septic shock patients, evaluated the impact of administering norepinephrine early or late on the clinical outcomes. The pooled data from these studies showed that patients receiving norepinephrine in the early stage had significantly lower mortality rates than those who received norepinephrine late (OR: 0.44; 95% CI: 0.35 – 0.55; p<0.00001) (Figure 3). Similarly, the data pooled from 4 studies showed that patients in the early group took significantly less time (in hours) to achieve the target MAP than those in the late group (MD: -1.17; 95% CI: -2.00 – -0.34; p=0.0006) (Figure 4). However, the length of ICU stay did not differ statistically between the two groups (MD: 0.55; 95% CI: -0.52 – 1.62; p=0.31) (Figure 5).

Figure 3. A forest plot comparing mortality rates between patients receiving early and late norepinephrine.

Figure 4. A forest plot comparing time taken to achieve target MAP between patients receiving early and late norepinephrine.

Figure 5. A forest plot comparing length of ICU stay between patients receiving early and late norepinephrine.

Does the timing of Vasopressin administration improve outcomes of septic shock?

Five studies, including 1043 patients, assessed the impact of vasopressin as an adjunct therapy for septic shock. Data pooled from two studies showed that the early addition of vasopressin to norepinephrine has no significant mortality benefit compared to monotherapy norepinephrine (OR: 0.87; 95% CI: 0.48 – 1.58; p=0.65) (Figure 6). Similarly, data from the study in which vasopressin was administered in combination with hydrocortisone showed that early administration has no significant mortality benefit compared to norepinephrine in combination with hydrocortisone (OR: 1.20; 95% CI: 0.79 – 1.82; p=0.39) (Figure 6). However, the pooled data from two studies showed that early initiation of vasopressin significantly reduces mortality of septic shock patients compared to late initiation (OR: 0.60; 95% CI: 0.41 – 0.90; p=0.01) (Figure 6).

Figure 6. Forest plot for mortality rate after early vasopressin initiation.

In addition, our statistical analyses show that early addition of vasopressin to norepinephrine is associated with a significant reduction in the time taken to achieve target MAP compared to monotherapy norepinephrine (MD: -3.15; 95% CI: -4.40 – -1.90; p<0.00001) (Figure 7). However, this early addition shows no significant difference compared with norepinephrine alone on the incidence of new-onset arrhythmias (OR: 1.31; 95% CI: 0.45 – 3.78; p=0.62) or ICU length of stay (MD: -0.45; 95% CI: -3.25 – 2.36; p=0.75) (Figures 8 & 9).

Figure 7. Forest plot for time taken to achieve target MAP after early vasopressin initiation.

Figure 8. Forest plot for incidence of new-onset arrhythmias after early vasopressin initiation.

Figure 9. Forest plot for length of ICU stay after early vasopressin initiation.

The statistical analyses also suggest that early vasopressin in combination with hydrocortisone shows no significant difference compared to norepinephrine on the incidence of new-onset arrhythmias (OR: 0.39; 95% CI: 0.08 – 2.04; p=0.27) (Figure 8) and length of ICU stay (MD: 0.50; 95% CI: -0.01 – 1.01; p=0.05) (Figure 9). Similarly, patients receiving early or late vasopressin have statistically equal length of ICU stays (MD: -0.41; 95% CI: -3.64 – 2.82; p = 0.80) (Figure 9) and incidences of new-onset arrhythmias (OR: 0.69; 95% CI: 0.18 – 2.58; p=0.58) (Figure 8).

Impact of norepinephrine timing on the clinical outcomes of septic shock

The SSC guidelines have recommended using norepinephrine as the first-line vasopressor for septic shock.^6,7 The goal of this therapy is usually to attain and maintain a MAP of above 65 mmHg.^31,32 MAP is a crucial measure of pressure because if it goes excessively low, it can cause organs not to be perfused, causing the cells to start dying and organs not function properly.⁷ Furthermore, MAP below 65mmHg results in an elevated mortality risk because of the inadequate amount of oxygen delivered to the organs.⁷ Therefore, it is essential to assess how the timing of norepinephrine initiation affects MAP. Our statistical analysis indicated that early initiation of norepinephrine significantly decreases the time to achieve target MAP. This finding collaborates with what was already reported in a previous meta-analysis of 5 studies. According to that study, the duration it took to attain target MAP was shorter with early administration of norepinephrine (MD: -1.39; 95% CI: -1.81 – -0.96; p<0.00001).³³ Similarly, Hamzaoui and colleagues found that among patients with severe shock, early norepinephrine (within 6 hours of septic shock) was effective in obtaining the target MAP.³⁴ This decrease in the duration to attain MAP goal is particularly important as studies have shown that MAP is a key determinant of patients’ outcomes.^35,36

The current research has also revealed that early shock treatment with norepinephrine results in a considerable reduction in mortality. In similar fashion, Li and colleagues discovered that short-term mortality was lower when norepinephrine was administered early as opposed to late (p<0.00001).³³ The survival benefit reported in these studies can be attributed to faster achievement of MAP with early norepinephrine administration. In contrast, a 2019 double-blinded RCT recorded an insignificant difference in crude 28-day mortality between early and late norepinephrine use (p=0.15).²¹ However, the mortality benefit of early norepinephrine cannot be inferred from this study due to various limitations. First, the trial was not designed to assess mortality. Secondly, the fluid resuscitation rate was not controlled, and this might have affected the mortality outcomes. Finally, the trial was performed in a single center, and its findings cannot be generalized to other care facilities.

Early administration of norepinephrine also displays other benefits over late administration. Research has shown that early norepinephrine can boost cardiac output via multiple methods. One of these methods is the elevation of cardiac preload and reduction of preload reliance^34,37 during the initial stages septic shock by elevating mean systemic filling pressure and dispersing blood out of the disproportionately bumped unstressed volume to the stressed volume via α-adrenergic-mediated reduction of venous capacitance.³⁸ Additionally, norepinephrine can elevate cardiac output by boosting the ventricular contractility.³⁹ Hamzaoui and colleagues found that early norepinephrine administration in septic shock patients receiving adequate intravenous fluid might enhance the left ventricular ejection fraction and other indicators of left and right systolic function.³⁹ This effect was attributed to an improvement in coronary perfusion pressure via an increased diastolic blood pressure and β-adrenergic stimulation of cardiomyocytes at the initial stages of septic shock when the β-adrenergic receptors are poorly controlled.

Early norepinephrine administration can also prevent fluid overload. Research has shown that positive fluid balance is independently associated with poor outcomes in septic shock.^40,41 Therefore, early initiation of norepinephrine can result in a reduced amount of infused fluids, thus lowering the risk of fluid overload. This hypothesis is evident in the study by Li and colleagues, where patients receiving early norepinephrine had significantly lower amounts of infused fluids within the first 24 hours than those receiving norepinephrine late (p=0.03).²⁸ Similarly, Bai and colleagues found that the volume of intravenous fluids at 24 hours was significantly lower in patients receiving norepinephrine within 2 hours of septic shock onset than those receiving the therapy after 2 hours (p<0.001).²⁷ In contrast, permpikul and colleagues found that the amount of infused resuscitation fluids did not differ between patients in the early and late norepinephrine groups.²¹ This insignificant difference can be explained by the fact the study was performed during the 2013 – 2017 period when the SSC guidelines were used, meaning that fluid was given towards a target intravascular volume and central venous pressure. Additionally, early initiation of norepinephrine in cases of severe hypotension might recruit microvessels and improve microcirculation by increasing organ perfusion pressure. This is evident in the study by Georger and colleagues, who found that early norepinephrine significantly improved the tissue muscle oxygen saturation alongside MAP in septic shock patients.⁴²

Although we have shown that early administration of norepinephrine may be beneficial in septic shock patients, it is not without consequences. One of the well-known adverse events associated with vasopressor therapy is the development of arrhythmias. Permpikul et al. reported that early norepinephrine was associated with significantly lower incidences of new-onset arrhythmias and cardiogenic pulmonary edema than late norepinephrine (p=0.03 and 0.004, respectively).²¹ However, no other study has collaborated this finding, meaning that further high-quality trials are required to investigate whether administering norepinephrine in the early phase of septic shock might reduce the occurrence of new-onset arrhythmias and other adverse events.

Impact of vasopressin timing on the clinical outcomes of septic shock

Vasopressin is an antidiuretic drug that attaches onto the arginine vasopressin receptors. Currently, this medication is recommended as a supplementary treatment option for septic shock patients who are incapable of attaining or maintaining a MAP goal of more than 65mmHg on norepinephrine alone. In our analysis, we observed that the prompt addition of vasopressin to norepinephrine resulted to a considerable decrease in the time taken to achieve target MAP than norepinephrine alone. This faster achievement of target MAP suggests that resolving vasopressin insufficiency sooner could lessen the duration spent on the initial stages of septic shock. Since prior research has revealed that delay in obtaining objective MAP can raise the overall mortality of septic shock patients, the findings of this meta-analysis might provide clinicians with a therapy that can be employed to attain and sustain the target MAP within a shorter period. However, subsequent studies should concentrate on a precision therapeutic strategy to identify patients more likely to benefit from the early addition of vasopressin.

The pooled analyses have also shown that despite early vasopressin reducing the time to target MAP, it has no significant difference to norepinephrine alone in terms of mortality, ICU stay, and incidence of new-onset arrhythmias. These findings are collaborated by a previous meta-analysis of 5 studies which found that early vasopressin did not improve short-term mortality (p=0.60), length of ICU stay (p=0.77), and incidence of new-onset arrhythmias (p=0.10).⁴³ Therefore, there is still a need for further large-scale trials to evaluate the benefit of early concomitant vasopressin and norepinephrine therapy in septic shock patients.

In addition, evidence suggests that early induction of vasopressin as a supplement to norepinephrine might aid in resolving organ dysfunction. A 2017 study noted that patients who were managed with concomitant early vasopressin and norepinephrine had considerable decreases in Sequential Organ Failure Assessment (SOFA) scores within 72 hours than those who received the initial norepinephrine monotherapy (-4 vs. -1; p=0.12, respectively).²⁴ However, a 2018 randomized trial did not demonstrate any substantial variations in SOFA scores among patients receiving early vasopressin combined with norepinephrine and those receiving norepinephrine alone (p=0.29).²³ The evidence in these trials has shown contradictions, meaning that the early initiation of vasopressin to norepinephrine has the potential to decrease morbidity and time taken to resolve septic shock phase by resolving organ failure, but this potential benefit should be investigated further in large high-quality randomized trials.

Our subgroup analysis has also shown that early initiation of vasopressin as an adjunct therapy for septic shock patients receiving catecholamines is associated with a significant decrease in mortality than late initiation. However, the analyses have indicated a high interstudy heterogeneity, which suggests variation in outcomes. Reardon et al.²⁰ found no significant difference in mortality rate between the early and late vasopressin groups. On the other hand, Rydz and colleagues found that the in-hospital mortality was significantly higher in patients receiving vasopressin after 7 hours.³⁰ The difference between these studies may be attributed to the fact that Rydz and colleagues administered vasopressin at lower doses of norepinephrine, which might have statistically reduced the composite outcome of in-hospital mortality. This hypothesis is supported by a previous study which found that for every 10 ug/min increase in norepinephrine dose equivalents, less than 60 ug/min at the time of vasopressin initiation was associated with an increase in mortality by 20.7%.⁴⁴

The present study has also demonstrated a negligible variation in the occurrence of new-onset arrhythmias between the early and delayed vasopressin groups. In contrast, Reardon and colleagues reported a significantly lower incidence of new-onset arrhythmias when vasopressin was administered in the initial stages of septic shock.²⁰ This finding may be attributed to the fact that most patients in the early group were younger than those in the late group and were not necessarily at equal risk for developing arrhythmias. However, the outcomes of this research might not be applied to other healthcare institutions since it was undertaken in only one institution. Therefore, more RCTs are required to validate these findings.

Although we did not assess the effect of vasopressin timing on organ dysfunction, it is worth noting the effect on acute kidney injury (AKI), which is a common complication of sepsis and septic shock. A previous meta-analysis focusing on the clinical outcomes of septic shock patients receiving early vasopressin found that patients receiving vasopressin had a reduced need for renal replacement therapy (RRT).⁴³ Similarly, Rydz and colleagues reported that the incidence of AKI and the need for RRT was significantly lower for patients receiving adjunct vasopressin within 7 hours of stroke onset.³⁰ On the other hand, a previous meta-analysis focusing on the renal outcomes of distributive shock patients found that vasopressin and its analogs is associated with reduced AKI incidences and the need for RRT.⁴⁵ Therefore, vasopressin administration seems to be beneficial in resolving organ dysfunction, especially AKI. However, further large-scale RCTs are still needed to validate this important finding.

Implications for clinical care and future research

From a practical standpoint, early administration of norepinephrine can rapidly increase MAP and lower the death rate of septic shock patients. Therefore, expedited norepinephrine infusion may be a logical therapy for improving organ perfusion, and we can recommend that norepinephrine is initiated concurrently with resuscitation fluids at the initial stages of septic shock. However, this technique ought not to hinder sufficient fluid resuscitation as clinicians should be aware of the ‘vasoconstrictor-masked hypovolemia’.⁴⁶ Timely infusion of vasopressin as an additional treatment in septic shock might decrease mortality. However, further randomized trials are needed to validate this finding. Furthermore, early introduction of vasopressin as a supplement to norepinephrine rapidly improves the achievement of target MAP than norepinephrine alone. Therefore, we recommend early concomitant administration of vasopressin and norepinephrine in patients who are unable to attain the goal MAP or those with presumed or proven deficit of vasopressin.

Our contemporary qualitative review has also identified gaps that should be addressed in future research. First, the definition of early vasopressor initiation varied across the studies. In studies on norepinephrine, early initiation was either referenced to the onset of septic shock or administration of loading fluid, while for studies on vasopressin, early initiation referred to the time between administration of first-line vasopressor to the addition of vasopressin. Therefore, future studies should look to harmonize the definition of early initiation to determine the optimal time at which vasopressor therapy can have improved outcomes. Secondly, evidence suggests that early initiation of norepinephrine might reduce the incidences of new-onset arrhythmias; however, further research is required to establish this finding. Finally, there is a suggestion that early administration of vasopressin might aid in resolving organ dysfunction; however, rigorous reporting is still required to validate this finding.

Limitations

The present review article has several limitations that should be accounted for when interpreting our findings. First, we observed high heterogeneity in some of the statistical analyses. This heterogeneity might be a result of varied sample sizes, different dosages, and methodological designs. However, the random effect model employed in our analyses was justified and countered the heterogeneity to provide conservative outcomes. Second, our eligibility criteria excluded articles of pediatric patients; hence, our findings can only be inferred for adult septic shock patients. Thirdly, due to limited data, we were unable to carry out a meta-analysis on organ dysfunction, which is an important clinical outcome in septic shock. However, we have addressed this outcome by reviewing the literature qualitatively. Fourth, our study was designed to evaluate the effect of various vasopressor timing on clinical outcomes; however, we only identified articles reporting the timing of norepinephrine and vasopressin. Finally, most of the studies in this review were non-randomized and conducted in single centers; hence, the bias that comes with such study designs was transferred to our analyses.